Ceramics International 33 (2007) 779784

www.elsevier.com/locate/ceramint

The effects of transition metals on carbothermal

synthesis of b-SiC powder
Yung-Jen Lin *, Chih-Ming Chuang
Department of Materials Engineering, Tatung University, 40 Chung-san North Road,
Section 3, 10451 Taipei, Taiwan
Received 8 August 2005; received in revised form 22 December 2005; accepted 30 January 2006
Available online 18 April 2006

Abstract
Several transition metals were added in the carbothermal syntheses of b-SiC powder from silica and phenolic resin at temperatures 1500 8C.
The effects of the transition metals on the yields of SiC and powder morphology were investigated. The results showed that by adding 1 wt.% of Fe,
Co or Ni in the silica/phenolic resin mixtures, the synthesizing temperature of SiC decreased and the yields of SiC increased. But adding Mn, Pd or
Cu has no favorable effects on the syntheses of b-SiC powder. After reaction at 1400 8C, the highest yield was obtained by adding 1 wt.% Fe, which
reaches 77%. Adding Co could result in 73% yield and adding Ni could only have 30% yield. The SiC powder obtained from mixtures with 1 wt.%
Fe was agglomerated, while that from mixtures with 1 wt.% Co was mostly un-agglomerated particles of 100300 nm in size. The positive effects
of transition metals in SiC powder synthesis at temperatures <1500 8C seems to be in the enhancement of SiO(g) formation.
# 2006 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Keywords: Combinatorial synthesis; Ceramic; Additives; Morphology; Powder

1. Introduction high purity b-SiC powder but also involves inexpensive

Silicon carbide (SiC) powder is an excellent reinforcement
for ceramic composites because of its high hardness, high
strength, chemical inertness and oxidation resistance. It has
been exploited in various ceramic composites to improve their
structural properties [110]. There are several common
synthesizing methods for silicon carbide, including direct
carbonization of Si metals [11], CVD from silane [12,13],
solgel [14] and carbothermal reduction of silicon dioxide
[15,16]. These methods have their own advantages and
disadvantages. For example, the silicon carbide powder made
from chemical vapor deposition and solgel method has high
purity and narrow particle size distribution. But the reactants
are unstable and toxic and the production cost is high. Direct
carbonization method is simple and cheap, but it will leave
significant amounts of unreacted silicon and impurities. So
far, the carbothermal reduction method is the most prevailing
method for SiC powder synthesis [1520]. It not only gives
* Corresponding author. Tel.: +886 2 25866040; fax: +886 2 25936897.
E-mail address: yjlin@ttu.edu.tw (Y.-J. Lin).
0272-8842/$32.00 # 2006 Elsevier Ltd and Techna Group S.r.l. All rights reserved.
doi:10.1016/j.ceramint.2006.01.010
starting materials (e.g., silicon dioxide, rice husk, graphite
and charcoal).
Although the carbothermal reduction method only requires
cheap reactants and could have good products, it is usually
performed at 1600 8C or above in order to have high yield of b-
SiC [15]. At higher temperature (1600 8C or higher), the
reaction is believed to be a SiO(g)CO(g) gasgas reaction,
favoring whisker formation [17,21]. Transition metals
(especially Fe, Co, Ni) have been shown to act as catalysts
in the gasgas reaction to form whisker via VLS (vapor
liquidsolid) mechanism [17]. At lower temperatures (1500 8C
or lower), the reaction is attributed to the SiO(g)C(s) gassolid
reaction, favoring powder formation. Because of low
temperatures, the yield of SiC is usually low. Whether or
not the transition metals have similar effects in the lower-
temperature carbothermal synthesis of SiC powder appears to
be relatively unclear.
In this research, we surveyed a number of transition metal
additives in terms of their effects on the synthesis of SiC
powder at lower temperatures (1500 8C or lower), where gas
solid reaction dominates. The yields of SiC and the morphology
of products were studied.
780 Y.-J. Lin, C.-M. Chuang / Ceramics International 33 (2007) 779784

2. Experiment

Ludox11 (colloidal silica, 40 wt.% suspension in water,

particle size 2024 nm) and phenolic-resin2 (resol type in
methanol, 78% solid content, 50% in solid (carbon) content
after carbonization) were used as the silica and carbon sources,
respectively. Both chemicals were mixed in ethanol (95%) with
a weight ratio of C/SiO2 equaled to 5/3 (corresponded to C/SiO2
molar ratio equaled 3/1) after carbonization of phenolic resin.
Salts of transition metals were then added into the solution and
mixing with a stirrer for 1 h. After mixing, the mixtures were
dried under IR lamps. The dried cakes were then ground into
powder for subsequent heat treatments. The transition metal salts
used were PdCl26H2O, Mn(CH3COO)26H2O, FeCl36H2O,
Co(NO3)26H2O, Ni(NO3)26H2O and CuSO45H2O. The
amounts of these additives were such that after high temperature
reduction, the transition metals were 0.5, 1 and 3 wt.% of the
solid (carbonized carbon + silica) mixtures.
Fig. 1. X-ray diffraction patterns of the SiO2/resin mixture, without additives,
Carbothermal synthesis of b-SiC powders was performed after reaction at different temperatures in Ar.
with two-step heat treatments in flowing Ar (100 ml/min) in
alumina crucibles. The heat treatment was first to maintain at
1000 8C for 2 h (with 10 8C/min heating rate) to carbonize the with Pd or Mn additions show no crystalline phases.
phenolic resin [22], and then heated to 1300, 1400 or 1500 8C Consequently, the addition of Pd and Mn has no effects in
and soaked for 4 h. promoting SiC formation at 1300 8C while the addition of Fe,
The products were analyzed with X-ray diffractometry Co, Ni or Cu lowers the formation temperature of SiC. It is also
(XRD) for phase identification (D-5000, Siemens, Karlsruhe, noted that cristobalite peaks are present in the patterns for the
Germany; Cu Ka radiation). Quantitative analyses of the samples with Ni or Cu addition, indicating Ni or Cu also
powder were performed using weight loss and XRD internal enhances the crystallization of cristobalite (from silica).
standard method [23] with CaF2 as the internal standard. The The effectiveness of Fe, Co, Ni, and Cu addition in the
product powder was first heated in air at 650 8C for 1 h to promotion of SiC formation can be also confirmed by the
remove the residual (unreacted) carbon. The weight loss after weight loss after reaction. The gassolid reaction of SiC
this treatment is the amount of residual carbon in the product. formation proceeds in two steps [15,20,24]:
Then, the powders were analyzed with XRD internal standard
method to obtain the weight percentage of SiC in the powders
after carbon removal. The balance would be the amount of Cs SiO2s ! SiOg COg (1.1)
cristobalite/silica (SiO2). The morphology of the powder
products was observed in a scanning electron microscope (JSM 2Cs SiOg ! SiCs COg (1.2)
5600, JEOL, Tokyo, Japan).

3. Results

Fig. 1 is the X-ray diffraction patterns of SiO2/resin mixtures

without additives after reaction between 1300 and 1500 8C. The
patterns reveal that no b-SiC was produced after reaction at
1300 8C. This sample contained only amorphous carbon and
silica, as indicated by the hump near two-theta equals 228. As
the reaction temperatures increased to 1400 8C, b-SiC started to
appear. The b-SiC became the major crystalline phase after
reaction at 1500 8C.
Fig. 2 shows the diffraction patterns of SiO2/resin mixtures
with 1 wt.% transition metal additives and reacted at 1300 8C
for 4 h. Peaks of b-SiC are present in the patterns for samples
with Fe, Co, Ni or Cu additions. The XRD patterns for samples

1
AS 40, Dupont, Wilmington, DE, USA. Fig. 2. X-ray diffraction patterns of the SiO2/resin mixture, with 1 wt.% of
2
PF-650 Chung-chun Plastic Co. Ltd., Hsinchu, Taiwan. different transitional metal additives, after reaction at 1300 8C for 4 h in Ar.
 Y.-J. Lin, C.-M. Chuang / Ceramics International 33 (2007) 779784 781

According to Eqs. (1.1) and (1.2), the SiO2 needs to be reduced

to gaseous SiO(g) in order to react with C to form SiC. Carbon
monoxide (CO) is a by-product in both equations and will
dissipate so that the reactions can continue. Therefore, the
weight loss could indicate that the reactions take place. More
weight loss implies higher degree of reaction. If the additives
had no effects in the reaction, the weight loss after reaction at
each temperature would be the same as that from synthesizing
mixtures without additives. On the other hand, if the additives
promote one or both steps of the reactions, the weight loss
would be more than that from mixtures without additives under
the same conditions. Hence, the weight loss could be used as an
indicator of the proceeding of the reactions and the effective-
ness of the additives.
Fig. 3 is the weight loss of the SiO2/resin mixtures with
1 wt.% of different additives after reaction between 1300 and
Fig. 4. Weight percentages of SiC, residual carbon and SiO2 in the reaction
1500 8C. Without additives, significant difference in weight products with 1 wt.% additives reaction at 1400 8C. Also included were samples
loss occurred between 1400 and 1500 8C, which is consistent without additives. NA, no additives, reaction at 1400 8C; 15NA, no additives,
with the temperature range of SiC formation revealed in reaction at 1500 8C.
XRD (see Fig. 1). In contrast, the differences in weight losses
are equally important between 1300 and 1400 8C and
between 1400 and 1500 8C in samples with Fe, Co, Ni weight percentages of each component (SiC, residual carbon
additives. This implies that in samples without additives, the and SiO2) in the products with additives after reaction at
reaction to form SiC is important only above 1400 8C, while 1400 8C. The compositions of the products of reaction from
in samples with Fe, Co, Ni, significant reaction starts from mixture without additives are also included for comparison.
1300 8C. From this figure, it appears that the addition of 1 wt.% of Fe and
It is noted in Fig. 3 that the sample with Cu addition showed Co resulted in a dramatic increase in SiC yields after reaction at
different weight loss behavior. Although at 1300 8C, the weight 1400 8C. Without additives, the samples had 19% of SiC at
loss was significant, between 1300 and 1500 8C it was not as 1400 8C and 32% of SiC at 1500 8C. When 1 wt.% of Fe or Co
high as shown by other additives in the same temperature range. were used, the yields of SiC at 1400 8C reached 77 and 73%,
This indicates that the reactions started at low temperature respectively. These yields were much higher than the yield
(<1300 8C), but did not proceed in the same way as for the obtained from mixture without additives. In contrast, the
other samples at higher temperature. As seen further in Fig. 4, addition of Ni had only moderate increase in the SiC yield at
the samples with Cu have much lower SiC yield than the 1400 8C (30%). More noticeably, the addition of Cu had
samples with Fe, Co or Ni additives. adverse effect on the yield of SiC formation (7% from sample
To investigate the yields of SiC after reaction at 1400 8C, with 1 wt.% Cu versus 19% from sample without additives)
quantitative analyses were performed. Fig. 4 compares the even though it lowered the formation temperature of SiC (see

Fig. 3. Weight changes of the SiO2/resin mixtures with 1 wt.% of different Fig. 5. X-ray diffraction patterns of the SiO2/resin mixtures with 0.53 wt.%
additives after reaction between 1300 and 1500 8C, for 4 h, in Ar. Co, after reaction at 1400 8C for 4 h in Ar.